Team MIT and Team Delta Sigma Selected as Finalists in the 16th LSI Design Contest 2013

On March 15, 2013, the MIT Team and the Delta Sigma Team were selected as finalists in the 16th LSI Design Contest 2013. And last Wednesday, March 27, 2013, the two teams from STEI departed for the University of Ryukyus, Okinawa, Japan to present their work in front of the jury. The list of team members is as follows:

MIT (Muhammads In Team):
1. Muhammad Firmansyah Kasim (13209041)
2. Fadhli Zakiy (13209053)
3. Muhammad Fahreza (13209015)

Work title: Algorithm, Real-Time Architecture and Implementation of Adaptive General Spectral Subtraction
Abstrak singkat: A common noise cancellation algorithm is Spectral Subtraction (SS) method. In this paper, we propose a new SS method: Adaptive General Spectral Subtraction (AGSS) method that can adapt to small change in noise. We implement AGSS into hardware-software co-design. The software is to control audio codec, while the hardware has a role to do the whole AGSS process. In the hardware implementation, we propose small and fast square root and divider blocks. Our hardware design utilizes 5086 logic elements or 299.2 in the 50 XOR normalization units and additional 2.18kB memory. Latency of the hardware is 2.48μs or 159.1 normalization units with 87.54MHz maximum frequency. Moreover, implementation of AGSS process in hardware also gives speed-up up to 13000 times compared to its software implementation. Hence, our design provides a high-quality, small, and fast noise cancellation design.
Link video: http://www.youtube.com/watch?v=9mij02XVyM4

Delta Sigma Team:
1. Anugerah Firdauzi (13209109)
2. Muhammad Arijal (13209097)
3. Kiki Wirianto (13209071)

Work title: Design and Implementation of Real Time Noise Canceling System Based on Spectral Subtraction Method
Abstrak singkat: Noise cancelling system is a kind of signal processing module which function is to remove the noise in a signal. In this paper, a digital signal noise cancelling is designed based on Spectral Subtraction Method which cancels the noise in frequency domain by estimating the noise energy spectral from a noisy input. That method is performed in frequency domain with polar vector number, thus a time to frequency domain and rectangular to polar transformer with each of their complement is needed to perform the noise cancelling operation. The noise cancelling system is expected to be able to perform a real-time noise cancellation, that a full-hardware system implementation is considered a pretty good choice for a high speed system even though it is very challenging to implement such complex system this way.
Link video: http://www.youtube.com/watch?v=l4GUn87sCBQ

Good luck! May it all go smoothly, and may you become the best team. May you bring honor to STEI ITB.